Method for analyzing and evaluating facial muscle status

Abstract

A method for analyzing and evaluating facial muscle status includes following steps: capturing user's face image through an image capturing unit of a face image analyzing apparatus after it is activated; analyzing the face image through an analyzing algorithm for obtaining multiple ideal muscle identifying points corresponding to five sense features of a face in the face image; identifying the face image through a fuzzy comparison algorithm and a training model for obtaining multiple actual muscle identifying points corresponding to actual muscle status of the face in the face image; evaluating each of the actual muscle identifying points and generating evaluated results based on the multiple ideal muscle identifying points in company with a pre-stored evaluation rule; and, displaying the multiple ideal muscle identifying points, the multiple actual muscle identifying points and the evaluated results on a display of the face image analyzing apparatus.

What is claimed is: 
     
         1 . A method for analyzing and evaluating facial muscle status, applied to a face image analyzing apparatus having a processor, an image capturing unit, and a display, comprising following steps of:
 a) taking a photo of a user by the image capturing unit;   b) retrieving a face image of the user from the photo;   c) executing an analyzing algorithm by the processor to analyze the face image for obtaining multiple ideal muscle identifying points from the face image, wherein the multiple ideal muscle identifying points are automatically generated based on five sense features on a face in the face image;   d) identifying the face image through a fuzzy comparison algorithm in company with a training model by the processor for obtaining multiple actual muscle identifying points from the face in the face image, wherein the multiple actual muscle identifying points are corresponding to an actual muscle status of the face;   e) evaluating each of the multiple actual muscle identifying points by the processor according to each of the multiple ideal muscle identifying points and an evaluation rule for generating an evaluated result;   f) overlapping and displaying the face image, the multiple ideal muscle identifying points, and the multiple actual muscle identifying points on the display; and   g) displaying the evaluated result on the display.   
     
     
         2 . The method in  claim 1 , wherein the multiple ideal muscle identifying points comprise a first ideal muscle identifying point corresponding to an ideal position of right side risorius muscle of the face, a second ideal muscle identifying point corresponding to an ideal position of left side risorius muscle of the face, a third ideal muscle identifying point corresponding to an ideal position of right side masticatory muscle of the face, and a fourth ideal muscle identifying point corresponding to an ideal position of left side masticatory muscle of the face, and the multiple actual muscle identifying points comprise a first actual muscle identifying point corresponding to an actual position of right side risorius muscle of the face, a second actual muscle identifying point corresponding to an actual position of left side risorius muscle of the face, a third actual muscle identifying point corresponding to an actual position of right side masticatory muscle of the face, and a fourth actual muscle identifying point corresponding to an actual position of left side masticatory risorius muscle of the face. 
     
     
         3 . The method in  claim 1 , wherein the training model records multiple reference images respectively labeled with multiple muscle identifying points thereon, the step d) is to perform a fuzzy comparison process to the face image and the multiple reference images of the training model by the fuzzy comparison algorithm for obtaining at least one of the reference images which is determined approximate to the face image, and sets the positions of the multiple actual muscle identifying points on the face of the face image according to the positions of the multiple muscle identifying points labeled on the obtained reference image. 
     
     
         4 . The method in  claim 2 , wherein the step c) comprises following steps of:
 c11) identifying the face of the face image for at least obtaining the positions of a right eye, a left eye, a nose, and lips;   c12) virtually connecting an inner corner of the right eye with an outer corner of the right eye for generating a first horizontal reference line;   c13) logically segmenting the first horizontal reference line into four equal parts;   c14) virtually generating a first vertical reference line perpendicular to the first horizontal reference line from a position corresponding to a first part out of the four equal parts which is close to the outer corner of the right eye;   c15) logically segmenting a height from a nose ala of the nose to a nose tip of the nose into five equal parts;   c16) regarding a point upon the first vertical reference line located at a position corresponding to a second part out of the five equal parts counted downward from the nose ala as the first ideal muscle identifying point;   c17) logically segmenting a height from the nose tip to a jaw of the face into three equal parts; and   c18) regarding a point upon the first vertical reference line located at a position corresponding to a first part out of the three equal parts counted downward from the nose tip as the third ideal muscle identifying point.   
     
     
         5 . The method in  claim 4 , wherein the step c11) is to identify the face image through Dlib Face Landmark system for obtaining the positions of the right eye, the left eye, the nose, and the lips. 
     
     
         6 . The method in  claim 2 , wherein the step c) comprises following steps of:
 c21) identifying the face of the face image for at least obtaining the positions of a right eye, a left eye, a nose, and lips;   c22) virtually connecting an inner corner of the left eye with an outer corner of the left eye for generating a second horizontal reference line;   c23) logically segmenting the second horizontal reference line into four equal parts;   c24) virtually generating a second vertical reference line perpendicular to the second horizontal reference line from a position corresponding to a first part out of the four equal parts which is close to the outer corner of the left eye;   c25) logically segmenting a height from a nose ala of the nose to a nose tip of the nose into five equal parts;   c26) regarding a point upon the second vertical reference line located at a position corresponding to a second part out of the five equal parts counted downward from the nose ala as the second ideal muscle identifying point;   c27) logically segmenting a height from the nose tip to a jaw of the face into three equal parts; and   c28) regarding a point upon the second vertical reference line located at a position corresponding to a first part out of the three equal parts counted downward from the nose tip as the fourth ideal muscle identifying point.   
     
     
         7 . The method in  claim 6 , wherein the step c21) is to identify the face image through Dlib Face Landmark system for obtaining the positions of the right eye, the left eye, the nose, and the lips. 
     
     
         8 . The method in  claim 2 , wherein in the step e), the processor evaluates the status of the first actual muscle identifying point and the second actual muscle identifying point as average if the distances between the positions of the first actual muscle identifying point and the second actual muscle identifying point and the positions of the first ideal muscle identifying point and the second ideal muscle identifying point are smaller than or equal to a threshold, evaluates the status of the first actual muscle identifying point and the second actual muscle identifying point as good or excellent if the positions of the first actual muscle identifying point and the second actual muscle identifying point are higher than the positions of the first ideal muscle identifying point and the second ideal muscle identifying point and the distances between the positions of the first actual muscle identifying point and the second actual muscle identifying point and the positions of the first ideal muscle identifying point and the second ideal muscle identifying point are bigger than the threshold, and evaluates the status of the first actual muscle identifying point and the second actual muscle identifying point as fair or poor if the positions of the first actual muscle identifying point and the second actual muscle identifying point are lower than the positions of the first ideal muscle identifying point and the second ideal muscle identifying point and the distances between the positions of the first actual muscle identifying point and the second actual muscle identifying point and the positions of the first ideal muscle identifying point and the second ideal muscle identifying point are bigger than the threshold. 
     
     
         9 . The method in  claim 2 , wherein in the step e), the processor evaluates the status of the third actual muscle identifying point and the fourth actual muscle identifying point as excellent if the distances between the positions of the third actual muscle identifying point and the fourth actual muscle identifying point and the positions of the third ideal muscle identifying point and the fourth ideal muscle identifying point are smaller than or equal to a threshold, and evaluates the status of the third actual muscle identifying point and the fourth actual muscle identifying point as average, fair, or poor in an order if the positions of the third actual muscle identifying point and the fourth actual muscle identifying point are lower than the positions of the third ideal muscle identifying point and the fourth ideal muscle identifying point and the distances between the positions of the third actual muscle identifying point and the fourth actual muscle identifying point and the positions of the third ideal muscle identifying point and the fourth ideal muscle identifying point are bigger than the threshold. 
     
     
         10 . The method in  claim 2 , wherein the evaluation rule is logically segmenting a height from the position of a nose tip to the position of a nose ala of the face in the face image into five equal parts, and evaluates the status of the first actual muscle identifying point and the second actual muscle identifying point as average if the distance between the positions of the first actual muscle identifying point and the second actual muscle identifying point and the positions of the first ideal muscle identifying point and the second ideal muscle identifying point are smaller than or equal to one part out of the five equal parts, evaluates the status of the first actual muscle identifying point and the second actual muscle identifying point as good if the positions of the first actual muscle identifying point and the second actual muscle identifying point are higher than the positions of the first ideal muscle identifying point and the second ideal muscle identifying point by one part out of the five equal parts, evaluates the status of the first actual muscle identifying point and the second actual muscle identifying point as excellent if the positions of the first actual muscle identifying point and the second actual muscle identifying point are higher than the positions of the first ideal muscle identifying point and the second ideal muscle identifying point by two parts out of the five equal parts, evaluates the status of the first actual muscle identifying point and the second actual muscle identifying point as fair if the positions of the first actual muscle identifying point and the second actual muscle identifying point are lower than the positions of the first ideal muscle identifying point and the second ideal muscle identifying point by one to two parts out of the five equal parts, and evaluates the status of the first actual muscle identifying point and the second actual muscle identifying point as poor if the positions of the first actual muscle identifying point and the second actual muscle identifying point are lower than the positions of the first ideal muscle identifying point and the second ideal muscle identifying point by three or more parts out of the five equal parts. 
     
     
         11 . The method in  claim 2 , wherein the evaluation rule is logically segmenting a height from the position of a nose tip to the position of a jaw of the face in the face image into fifteen equal parts, and evaluates the status of the third actual muscle identifying point and the fourth actual muscle identifying point as excellent if the distances between the positions of the third actual muscle identifying point and the fourth actual muscle identifying point and the positions of the third ideal muscle identifying point and the fourth ideal muscle identifying point are smaller than or equal to one part out of the fifteen equal parts, evaluates the status of the third actual muscle identifying point and the fourth actual muscle identifying point as good if the positions of the third actual muscle identifying point and the fourth actual muscle identifying point are lower than the positions of the third ideal muscle identifying point and the fourth ideal muscle identifying point by at least one part out of the fifteen equal parts, evaluates the status of the third actual muscle identifying point and the fourth actual muscle identifying point as average if the positions of the third actual muscle identifying point and the fourth actual muscle identifying point are lower than the positions of the third ideal muscle identifying point and the fourth ideal muscle identifying point by at least two parts out of the fifteen equal parts, evaluates the status of the third actual muscle identifying point and the fourth actual muscle identifying point as fair if the positions of the third actual muscle identifying point and the fourth actual muscle identifying point are lower than the positions of the third ideal muscle identifying point and the fourth ideal muscle identifying point by three to four parts out of the fifteen equal parts, and evaluates the status of the third actual muscle identifying point and the fourth actual muscle identifying point as poor if the positions of the third actual muscle identifying point and the fourth actual muscle identifying point are lower than the positions of the third ideal muscle identifying point and the fourth ideal muscle identifying point by five or more parts out of the fifteen equal parts. 
     
     
         12 . The method in  claim 2 , wherein the evaluated result comprises textual descriptions or scores of each of the multiple actual muscle identifying points, or ratio of each of the multiple actual muscle identifying points in comparison with each of the multiple ideal muscle identifying points. 
     
     
         13 . The method in  claim 2 , further comprising following steps:
 h1) simulating a first virtual triangle constituted by the first actual muscle identifying point, a position of nasion of the face in the face image, and a position of right side temple of the face in the face image by the processor;   h2) simulating a second virtual triangle constituted by the second actual muscle identifying point, the position of nasion of the face, and a position of left side temple of the face in the face image by the processor;   h3) simulating a third virtual triangle constituted by the third actual muscle identifying point, a position of nose tip of the face in the face image, and a position of right side cheek of the face in the face image by the processor;   h4) simulating a fourth virtual triangle constituted by the fourth actual muscle identifying point, the position of nose tip of the face, and a position of left side cheek of the face in the face image by the processor;   wherein, square measures of the first virtual triangle, the second virtual triangle, the third virtual triangle, and the fourth virtual triangle are inversely proportional to the muscle status of the user.